Bending Spoons and the economics of AI-run software rollups

Bending Spoons, an Italian software company most people have never heard of, began trading on Nasdaq on 1 July 2026 at roughly an $18B valuation. While the market worries that AI will make software worthless, Bending Spoons is betting that AI makes existing, well-loved software cheap enough to buy and run at scale. Its own prospectus shows why. It runs less like a software company than like a private equity firm that never sells. Over the past decade it has bought 50+ products, including Evernote, WeTransfer, Vimeo, Eventbrite and now AOL, and folded them into one engineering org in Milan that shares infrastructure and payment systems across the whole portfolio. The playbook is blunt: buy a beloved app whose growth has stalled, cut costs hard, improve the product, raise prices. Evernote's personal plan went up 63%. StreamYard's base tier went up 80%. The part that matters if you build with AI is in the prospectus. Bending Spoons says AI authored or co-authored more than 90% of its pull requests by early 2026, up from under 10% a year earlier, with about 70% written by AI alone. That's pull requests, not lines of code, and it's still a huge shift in twelve months. It's what makes the math work. When the cost of maintaining and improving 50 products drops that far, you can afford to keep what everyone else flips. Revenue grew 95% in 2025 to $1.31B. Most of that came from acquisitions; organic growth was around 13%. For a rollup, acquisition-led growth is the thesis, so I don't read that as a weakness. The timing is deliberate. The "SaaSpocalypse" left a generation of pre-AI software cheap and unloved, because investors assumed AI would let anyone rebuild it over a weekend. Bending Spoons read it the other way. A loyal, sticky audience is hard to rebuild, and the same AI everyone is afraid of lets you run those products for a fraction of the old cost. As a builder, I find this the most interesting use of cheap AI-written code I've seen at scale. Most AI coding talk is about shipping new things faster. Bending Spoons applies the same falling costs to maintenance, and the long tail of software that already has users may be the bigger prize. From the investor's side, falling creation costs hurt incumbents, and they also put incumbents on the shopping list. Bending Spoons' prospectus points to 1,000+ more businesses that fit its profile. If the model holds, more of the software market could end up as AI-run rollups of apps people already use.